floyo logo
Powered by
ThinkDiffusion
Wan 2.6 is now live. Check it out 👉🏼
floyo logo
Powered by
ThinkDiffusion
Wan 2.6 is now live. Check it out 👉🏼

LTX 2.3 High Quality Production Ready Output

LTX 2.3 22B — high quality output, no setup required. Load your image, drop in your prompt, hit run. This workflow uses the full 22B model with settings tuned for cinematic, photorealistic results.

16

Generates in about -- secs

Nodes & Models

PrimitiveFloat
GetNode
ManualSigmas
KSamplerSelect
RandomNoise
UNETLoader
ltx-2.3-22b-dev.safetensors
LatentUpscaleModelLoader
ltx-2.3-spatial-upscaler-x2-1.0.safetensors
VAELoader
LTX23_video_vae_bf16.safetensors
VAELoaderKJ
LTX23_audio_vae_bf16.safetensors
DualCLIPLoader
gemma_3_12B_it.safetensors
ltx\ltx-2.3_text_projection_bf16.safetensors
LoadImage
INTConstant
SetNode
LTXVCropGuides
CFGGuider
EmptyImage
VAEDecode
LTXVScheduler
LTXVImgToVideoInplace
ResizeImageMaskNode
CLIPTextEncode
LoraLoaderModelOnly
ltx-2-19b-lora-camera-control-dolly-in.safetensors
ltx-2.3-22b-distilled-lora-384.safetensors
SimpleCalculatorKJ
LTXVLatentUpsampler
ImageScaleBy
LTXVEmptyLatentAudio
SamplerCustomAdvanced
LTXVConcatAVLatent
LTXVConditioning
LTX2LoraLoaderAdvanced
ltx-2-19b-ic-lora-detailer.safetensors
GetImageSize
LTXVSeparateAVLatent
LTXVPreprocess
EmptyLTXVLatentVideo
LTXVImgToVideoInplaceKJ
LTXVAudioVAEDecode
CR Float To Integer
VHS_VideoCombine
VHS_VideoCombine
easy showAnything

This workflow is set up for high quality LTX 2.3 output using the full 22B model — no configuration needed. Load your image, add your prompt, and run. Settings are pre-tuned for cinematic, photorealistic results.

Built for portraits, talking heads, slow atmospheric motion, and social verticals. If you need complex action or fast movement, Wan 2.5 will serve you better — this is for shots that need to look like real camera footage.

Read more

N